Image created with Flux Pro v1.1 Ultra. Image prompt: Meta, continuous infinity loop formed by a smooth chain of small bananas, soft horizon reflection, photorealistic, editorial, minimal, high detail, 3:2 landscape
Exclusive: Meta created flirty chatbots of Taylor Swift, other celebrities without permission | Reuters https://www.reuters.com/business/meta-created-flirty-chatbots-taylor-swift-other-celebrities-without-permission-2025-08-29/
More details emerge of rocky start to Meta Superintelligence Labs – Sherwood News https://sherwood.news/tech/more-details-emerge-of-rocky-start-to-meta-superintelligence-labs/
Meta’s AI leaders discuss using Google, OpenAI models in apps, The Information says | Reuters https://www.reuters.com/business/metas-ai-leaders-discuss-using-google-openai-models-apps-information-says-2025-08-30/
Apple loses 4 top AI researchers as key robotics lead heads to Meta, others join OpenAI and Anthropic – India Today https://www.indiatoday.in/technology/news/story/apple-loses-4-top-ai-researchers-as-key-robotics-lead-heads-to-meta-others-join-openai-and-anthropic-2781133-2025-09-03
For llama.vim the recommended setup now is Qwen 3 Coder 30B A3B Instruct: brew install llama.cpp llama-server –fim-qwen-30b-default Amazingly, on Macs the 30B MoE model performs better than the old Qwen 2.5 Coder 7B so if you have the necessary RAM it’s better to switch to https://x.com/ggerganov/status/1961471397428883882
@SemiAnalysis_ The problem is there’s tons of tests in different subsystems that get skipped in PyTorch. At Meta we rely on many devs, oncalls, heroics and BE weeks where we unskip or fix flaky tests. We love Jeff! And we need more people like him to build expertise in all pytorch subsystems”” / X https://x.com/marksaroufim/status/1963844930620600457
Meta introduces Set Block Decoding (SBD), a new inference accelerator for LLMs SBD samples multiple future tokens in parallel, cuts forward passes by 3–5x, needs no arch changes, stays KV-cache compatible, and matches NTP training performance. https://x.com/arankomatsuzaki/status/1963817987506643350
Hermes 4: Nous Research Open-Weight Reasoning Family Models – 70B / 405B (Llama-3.1 bases, released) – 14B (Qwen3 base, research baseline) Hermes 4 70B & 405B – Base: Llama-3.1-70B / 405B – Training: TorchTitan (modified), Axolotl, 192× B200s, FSDP and TP – Dataset: 56B tokens https://x.com/gm8xx8/status/1962943078702186627
Goated FAIR team just found how coding agents sometimes “”cheat”” on SWE-Bench Verified. It’s really simple. For example, Qwen3 literally greps all commit logs for the issue number of the issue it needs to fix. lol, clever model. “”cheat”” cuz it’s more like env hacking. https://x.com/giffmana/status/1963327672827687316
